arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2008.05247cs.LGcs.CVstat.ML
收藏

学习从错误中学习:对抗性噪声的鲁棒优化

Learning to Learn from Mistakes: Robust Optimization for Adversarial Noise

Radboud University(拉德堡德大学);Leiden University(莱顿大学)

Alex Serban, Erik Poll, Joost Visser

摘要:

对对抗性噪声的敏感性阻碍了机器学习算法在安全关键应用中的部署。虽然已经提出了许多对抗性防御,但对抗性噪声的鲁棒性仍然是一个悬而未决的问题。最引人注目的防御,对抗训练,需要大量增加处理时间,它已被证明是过拟合的训练数据。在本文中,我们的目标是通过在低数据状态下训练鲁棒模型并在不同模型之间转移对抗性知识来克服这些限制。我们训练一个元优化器,它学习使用对抗性示例来鲁棒地优化模型,并且能够将学到的知识转移到新模型中,而不需要生成新的对抗性示例。实验结果表明,元优化器在不同的架构和数据集上是一致的,这表明可以自动修补对抗性漏洞。

英文摘要:

Sensitivity to adversarial noise hinders deployment of machine learning algorithms in security-critical applications. Although many adversarial defenses have been proposed, robustness to adversarial noise remains an open problem. The most compelling defense, adversarial training, requires a substantial increase in processing time and it has been shown to overfit on the training data. In this paper, we aim to overcome these limitations by training robust models in low data regimes and transfer adversarial knowledge between different models. We train a meta-optimizer which learns to robustly optimize a model using adversarial examples and is able to transfer the knowledge learned to new models, without the need to generate new adversarial examples. Experimental results show the meta-optimizer is consistent across different architectures and data sets, suggesting it is possible to automatically patch adversarial vulnerabilities.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑